To ensure adequate inventory for upcoming builds, a necessity for reducing lead times, a procurement plan must be able to forecast demand. This involves a lot of data analysis and predictions that make it difficult to nail down an exact number, as unexpected fluctuations in market conditions can affect demand. However, with careful planning and room for error, you can gain insight into expected demand for coming months. There are a few different methods you can use to get a clear picture of how many orders you will likely get in the near future.
Factors to consider
Before the number crunching begins, you must first gather information. You will need records of order volumes from previous months to determine past demand. It's best if you can look at recent months as well as annual sales variances if possible, as this can help you define annual and seasonal fluctuations. The lead time of the product you're making as well as that of the various raw materials and components will impact the demand forecast as well. Another aspect that should not be overlooked is marketing planning, according to Indiana University's Kelley School of Business. If your company is planning to increase marketing efforts or perhaps offer a discount or create a sales promotion, you may see an unusual uptick in orders. It is important to look outside the company also. Economic conditions and competitor actions all influence demand as well.
This is perhaps the most straightforward method of determining the forecasted demand. According to SME Toolkit, all you have to do is add up the actual demand for the previous three months and divide the total by four (to account for the upcoming month). When "F" represents the forecast, "D" stands for demand and the number correlates with the month, the formula looks like:
F4 = (D1 + D2 + D3) ÷ 4
The above equation is great if you tend to have relatively steady demand, but if it fluctuates at all, you can easily adapt this to take the variation into consideration. The source suggested determining the weighted constant, which is typically a number between one and 10. The weight will be greater for more recent data, because the older the data, the less influence it has over the forecast. In this case, you'll want to determine the weighted moving average (WMA) by adding the sum of each month (D) as it is weighted (W) – and again the number is only there to represent the month:
WMA 4 = (W * D1) + (W * D2) +(W * D3)
The idea behind this method, which is a form of the weighted moving average model, is to determine the demand by not just establishing averages but also making them appear more fluid so it is easier to identify trends based on recent changes. SME Toolkit explained that this method can help eliminate random fluctuations that are less likely to affect the forecast for more accurate predictions. To determine the forecast (Ft)with exponential smoothing, the North Carolina State University Poole College of Management explained that it can boil down to three figures – the prior period's forecast (Ft-1), the actual demand for that period (At-1) and the weight being assigned to it (W) which ranges between zero and one. There are two ways to determine the forecast:
Ft = Ft-1 + W * (At-1 – Ft-1)
Ft = W * At-1 + (1 – W) * Ft-1
Moving averages and exponential smoothing are just two methods of forecasting demand, but there are many more that take into account a slew of data, such as the factors mentioned earlier, to hone down the predictions for better accuracy.